Today I want to touch on a little understood aspect of safe withdrawal rates (SWRs) in retirement. Using the historical method to determine SWRs (as opposed to Monte Carlo analysis) means that the maximum SWR (approx 4%) is singularly determined by one 30 year period, the 30 year retirement period for the 1966 retiree. The 1929 Great Depression retiree had it better than the 1966 retiree did. By looking at the details of the 1966 retiree’s situation we can gain some insight into implications for the more modern retirees and use it as a benchmark for more modern portfolio allocations.

What was so bad about the 30 year retirement period starting in 1966 is that it was the worst time in history to retire and solely defines the SWR for a 60/40 stock bond portfolio consisting of the SP500 and US 10 yr Government bonds (basically the benchmark retirement portfolio). At first glance the 1966 retiree did pretty well. The 30 yr return for a 60/40 portfolio starting in 1966 was 9.67% per year. Stocks returned 10.62% per year during this period and bonds returned 7.61% per year. These numbers often are used to scare modern retirees. You’ve seen the stories – future returns will not be what they were in the past due to current valuations and low yields and thus the 4% rule is dead. As I’ve written about before, maybe. But there is more insight here under the surface. What matters more in determining SWRs are real returns, inflation, and the sequence of returns. Lets look at each of these for the 1966 retiree.

The 9.67% annual return for the 60/40 portfolio starting in 1966 becomes a 4.46% real return after inflation, 5.41% for stocks and 2.39% for bonds. Now we’re starting to get a glimpse of the problem for the 1966 retiree. Low real returns due to high inflation. And inflation has a dual impact. Not only does it eat into returns but it also increases withdrawal rates to keep the standard of living constant. Inflation for the 30 year period for the the 1966 retiree was 5.21% per year, one of the highest on record. The final nail in the coffin for the 1966 retiree was the sequence of returns. It turns out that successful SWRs are more correlated with the initial 15 yr return for a retiree than for the full 30 year period return. Low or even negative returns early in retirement are the worst thing that can happen during your retirement. If we look at the first 15 years of the 1966 retiree’s returns the picture becomes much clearer. For the first 15 years, the 60/40 portfolio had a real return of 5.74% nominal but -0.18% per year after inflation. Stocks returned 0.8% per year during this period and bonds returned -2.60% per year. And to round it out inflation averaged 5.92% during this 15 year period. Brutal is all I can say. Take a look at the 1966 retiree’s portfolio value, withdrawals, and current withdrawal rates (CWR) during his/her 30 years.

And yet, despite the poor real returns, high inflation, and a terrible sequence of returns, the 4% rule worked. Which I think shows the robustness of the 4% rule. In order for the 4% SWR to fail in the future the situation needs to be worse than it was for the 1966 retiree; poor real returns, high inflation, and really bad early years. At the same time we have more tools and knowledge available to the investor today to combat such a possibility. For example, broader more diverse asset classes, alternative withdrawal strategies, and/or portfolio strategies that mitigate downside risk. All can be used to increase the likelihood of the future success of the 4% SWR.

Finally, another important insight here is that when looking at some of the modern portfolio allocation strategies like the IVY portfolio we need to take the results with a grain of salt. Most of the modern portfolios only have data going back to 1973. Well, if you only look at data from 1973 onwards you’ll miss the worst case period in history for retirement. For example, if you look at the SWRs for the various portfolios I track you’ll see SWRs way higher than 4%. But since the data for those portfolios doesn’t encompass 1966 they need to be adjusted downwards as I mentioned in the post.

In short, when thinking about future SWRs or looking at alternative asset allocation models or portfolio strategies as applied to retirement make sure you insist on 1966.

*P.S. All the data in this post is from my own calculations using historical returns for the SP500, the US 10 year government bond, and US Inflation using CPI.*

*Full Disclaimer - Nothing on this site should ever be considered advice, research or the invitation to buy or sell securities. These are my personal opinions only.*

Hi Paul. I may have missed it, but have you ever shared the formula you use to determine SWR?

Looks like the lighthouse is an amazing place to be this time of year. Hope to join you guys out there someday soon.

Thanks.

Hi John, its a relatively simple calculation (its basically a finite series). You setup an initial portfolio value and assume an initial SWR which determines the initial withdrawal. Each year you change the portfolio value by the investment return for that year, adjust the portfolio withdrawal for inflation, and subtract it from the portfolio value. You do this each year for 30 years. Then you go back and adjust the initial SWR until the portfolio value in year 30 is zero. Then you need to do this for each retiree in the study period. So you start with the 1929 retiree, then do it for the 1930 retiree, etc…..until the final 30 year period which started in 1982. The first portfolio among all the retirees that goes to zero determines the SWR for a 100% success rate.

So, simple calculation but tedious in practice. Its a huge spreadsheet.

Paul

Good timing Paul, I have been thinking about SWR a lot recently. I am slightly confused by your statement: “And inflation has a dual impact. Not only does it eat into returns but it also increases withdrawal rates to keep the standard of living constant.” It eats into returns because real returns are nominal returns minus inflation? But real returns account for inflation, so can’t you can keep the withdrawal rate the same if they are in terms of real returns? Alternatively, you could increase the withdrawal rate each year to match inflation, but then use nominal returns. Isn’t it overkill to use real returns AND increase the withdrawal rate by inflation?

Thanks,

Jeff

Hey Jeff, the returns used to calculate SWR are nominal returns. But as you say nominal returns are composed of real returns + inflation. So, inflation hits the return component as well as the withdrawal component. Think about a person who’s not retired and not making withdrawals. They are still impacted by inflation. Think about the opposite – a retired person with no investments. They are impacted by inflation via cost of living increases. If you have investments AND are trying to maintain your standard of living you get the double whammy.

Paul